Gui-Jun Jiang1,2, Ru-Kai Gao3, Min Wang4, Tu-Xiu Xie5, Li-Ying Zhan2, Jie Wei1, Sheng-Nan Sun6, Pei-Yu Ji7, Ding-Yu Tan7, Jing-Jun Lyu8. 1. Emergency Department, Renmin Hospital of Wuhan University, Wuhan, 430060, China. 2. Department of Critical Care Medicine, Renmin Hospital of Wuhan University, Wuhan, 430060, China. 3. Wuhan Britain-China School, Wuhan, 430022, China. 4. Medical College of Wuhan University of Science and Technology, Wuhan, 430065, China. 5. Department of General Practice, Renmin Hospital of Wuhan University, Wuhan, 430060, China. 6. Emergency Department, Qilu Hospital of Shandong University (Qingdao), Qingdao, 266000, China. 7. Emergency Department, Clinical Medical College of Yangzhou University, Northern Jiangsu People's Hospital, Yangzhou, 225001, China. 8. Emergency Department, Renmin Hospital of Wuhan University, Wuhan, 430060, China. lvjingjun@whu.edu.cn.
Abstract
OBJECTIVE: To examine the independent risk factors of type-2 myocardial infarction (T2MI) elicited by acute upper gastrointestinal bleeding (AUGIB), and to establish a nomogram model for the prediction of AUGIB-induced T2MI. METHODS: A nomogram model was established on the basis of a retrospective study that involved 533 patients who suffered from AUGIB in the Department of Critical Care Medicine (CCM) or Emergency Intensive Care Unit (EICU) of Renmin Hospital of Wuhan University, Wuhan, China, from January 2017 to December 2020. The predictive accuracy and discriminative power of the nomogram were initially evaluated by internal validation, which involved drawing the receiver operating characteristic (ROC) curve, calculating the area under the curve (AUC), plotting the calibration curve derived from 1000 resampled bootstrap data sets, and computing the root mean square error (RMSE). The predictive ability of the nomogram was further validated through the prospective and multicenter study conducted by the investigators, which enrolled 240 AUGIB patients [including 88 cases from Renmin Hospital of Wuhan University, 73 cases from Qilu Hospital of Shandong University (Qingdao), and 79 cases from Northern Jiangsu People's Hospital)], who were admitted to the Department of CCM or EICU, from February 2021 to July 2021. RESULTS: Among the 533 patients in the training cohort, 78 (14.6%) patients were assigned to the T2MI group and 455 (85.4%) patients were assigned to the non-T2MI group. The multivariate analysis revealed that age >65, hemorrhagic shock, cerebral stroke, heart failure, chronic kidney disease, increased blood urea nitrogen, decreased hematocrit, and elevated D-Dimer were independent risk factors for AUGIB-induced T2MI. All these factors were incorporated into the nomogram model. The AUC for the nomogram for predicting T2MI was 0.829 (95% CI, 0.783-0.875) in the internal validation cohort and 0.848 (95% CI, 0.794-0.902) in the external validation cohort. The calibration curve for the risk of T2MI exhibited good consistency between the prediction by the nomogram and the actual clinical observation in both the internal validation (RMSE=0.016) and external validation (RMSE=0.020). CONCLUSION: The nomogram was proven to be a useful tool for the risk stratification of T2MI in AUGIB patients, and is helpful for the early identification of AUGIB patients who are prone to T2MI for early intervention, especially in emergency departments and intensive care units.
OBJECTIVE: To examine the independent risk factors of type-2 myocardial infarction (T2MI) elicited by acute upper gastrointestinal bleeding (AUGIB), and to establish a nomogram model for the prediction of AUGIB-induced T2MI. METHODS: A nomogram model was established on the basis of a retrospective study that involved 533 patients who suffered from AUGIB in the Department of Critical Care Medicine (CCM) or Emergency Intensive Care Unit (EICU) of Renmin Hospital of Wuhan University, Wuhan, China, from January 2017 to December 2020. The predictive accuracy and discriminative power of the nomogram were initially evaluated by internal validation, which involved drawing the receiver operating characteristic (ROC) curve, calculating the area under the curve (AUC), plotting the calibration curve derived from 1000 resampled bootstrap data sets, and computing the root mean square error (RMSE). The predictive ability of the nomogram was further validated through the prospective and multicenter study conducted by the investigators, which enrolled 240 AUGIB patients [including 88 cases from Renmin Hospital of Wuhan University, 73 cases from Qilu Hospital of Shandong University (Qingdao), and 79 cases from Northern Jiangsu People's Hospital)], who were admitted to the Department of CCM or EICU, from February 2021 to July 2021. RESULTS: Among the 533 patients in the training cohort, 78 (14.6%) patients were assigned to the T2MI group and 455 (85.4%) patients were assigned to the non-T2MI group. The multivariate analysis revealed that age >65, hemorrhagic shock, cerebral stroke, heart failure, chronic kidney disease, increased blood urea nitrogen, decreased hematocrit, and elevated D-Dimer were independent risk factors for AUGIB-induced T2MI. All these factors were incorporated into the nomogram model. The AUC for the nomogram for predicting T2MI was 0.829 (95% CI, 0.783-0.875) in the internal validation cohort and 0.848 (95% CI, 0.794-0.902) in the external validation cohort. The calibration curve for the risk of T2MI exhibited good consistency between the prediction by the nomogram and the actual clinical observation in both the internal validation (RMSE=0.016) and external validation (RMSE=0.020). CONCLUSION: The nomogram was proven to be a useful tool for the risk stratification of T2MI in AUGIB patients, and is helpful for the early identification of AUGIB patients who are prone to T2MI for early intervention, especially in emergency departments and intensive care units.